Description of the bug
The DADA2 read tracking table in the HTML summary report mislabeled rows for samples whose names share a common prefix (AND where one of those samples has the prefix as the full sample name). In my case: EQUIP_BLANK, EQUIP_BLANK_3_25_2026, and EQUIP_BLANK_3_30_2026. The denoised/merged/nonchim columns were shuffled across these rows, causing two samples to show denoisedF > input (impossible?). However when I checked the ASV_table.csv it all became clear (the actual ASV table output was correct), the bug seems isolated to the HTML report tracking table.
This did NOT occur for my samples that share a prefix, I think this is particular to these three because some of them are a complete prefix of another sample name (EQUIP_BLANK). Maybe this is just a dumb negative sample naming scheme on my part. But thought I'd let you know regardless!
DADA2 read tracking table (from HTML report):
Sample | input | filtered | denoisedF | denoisedR | merged | nonchim
EQUIP_BLANK | 52,350 | 49,764 | 347 | 340 | 234 | 234
EQUIP_BLANK_3_25_2026 | 40,147 | 28,011 | 43,582 | 43,576 | 42,996 | 42,996
EQUIP_BLANK_3_30_2026 | 7,302 | 4,132 | 24,572 | 19,361 | 478 | 478
ASV_table.tsv total reads per sample (ground truth):
Sample | Total reads
EQUIP_BLANK | 42,996
EQUIP_BLANK_3_25_2026 | 478
EQUIP_BLANK_3_30_2026 | 234
VERSIONS:
nf-core/ampliseq: 2.15.0
MultiQC: 1.29
Nexflow: 24.10.6.5937
Used on local machine (WSL2)
Command used and terminal output
Relevant files
No response
System information
No response
Description of the bug
The DADA2 read tracking table in the HTML summary report mislabeled rows for samples whose names share a common prefix (AND where one of those samples has the prefix as the full sample name). In my case: EQUIP_BLANK, EQUIP_BLANK_3_25_2026, and EQUIP_BLANK_3_30_2026. The denoised/merged/nonchim columns were shuffled across these rows, causing two samples to show denoisedF > input (impossible?). However when I checked the ASV_table.csv it all became clear (the actual ASV table output was correct), the bug seems isolated to the HTML report tracking table.
This did NOT occur for my samples that share a prefix, I think this is particular to these three because some of them are a complete prefix of another sample name (EQUIP_BLANK). Maybe this is just a dumb negative sample naming scheme on my part. But thought I'd let you know regardless!
DADA2 read tracking table (from HTML report):
EQUIP_BLANK | 52,350 | 49,764 | 347 | 340 | 234 | 234
EQUIP_BLANK_3_25_2026 | 40,147 | 28,011 | 43,582 | 43,576 | 42,996 | 42,996
EQUIP_BLANK_3_30_2026 | 7,302 | 4,132 | 24,572 | 19,361 | 478 | 478
ASV_table.tsv total reads per sample (ground truth):
EQUIP_BLANK | 42,996
EQUIP_BLANK_3_25_2026 | 478
EQUIP_BLANK_3_30_2026 | 234
VERSIONS:
nf-core/ampliseq: 2.15.0
MultiQC: 1.29
Nexflow: 24.10.6.5937
Used on local machine (WSL2)
Command used and terminal output
Relevant files
No response
System information
No response